Prediction of biological activity

In the following sections we propose typical methods of unsupervised learning and pattern recognition, the aim of which is to detect patterns in chemical, physicochemical and biological data, rather than to make predictions of biological activity. These inductive methods are useful in generating hypotheses and models which are to be verified (or falsified) by statistical inference. Cluster analysis has... 

While principal components models are used mostly in an unsupervised or exploratory mode, models based on canonical variates are often applied in a supervisory way for the prediction of biological activities from chemical, physicochemical or other biological parameters. In this section we discuss briefly the methods of linear discriminant analysis (LDA) and canonical correlation analysis (CCA). Although there has been an early awareness of these methods in QSAR [7,50], they have not been widely accepted. More recently they have been superseded by the successful introduction of partial least squares analysis (PLS) in QSAR. Nevertheless, the early pattern recognition techniques have prepared the minds for the introduction of modem chemometric approaches. 

Weinstein, H., R. Osman, J. P. Green, and S. Topiol. 1981a. Electrostatic Potentials as Descriptors of Molecular Reactivity The Basis for Some Successful Predictions of Biological Activity. In Chemical Applications of Atomic and Molecular Electrostatic Potentials. P. Politzer and D. G. Truhlar, Eds. Plenum Press, New York. 

Some biologically important o-quinones can react with the superoxide ion giving catechol derivatives, which may play a role in many diseases. For example, compounds bearing a nitro-catechol moiety have been claimed to be efficient catechol-0-methyl transferase inhibitors (Suzuki et al. 1992, Perez et al. 1992). The transferase is the first enzyme in the metabolism of catecholamine a hyperactivity of this enzyme leads to Parkinson s disease. Therefore, prediction of biological activity and antioxidant properties of quinones is an important challenge for researchers. 

Stepanchikova AV, Lagunin AA, Filimonov DA, Poroikov W. (2003) Prediction of biological activity spectra for substances Evaluation on the diverse sets of drug-like structures. Curr. Med. Chem. 10 225-233. 

Harper G, Bradshaw J, Gittins JC, Green DVS, Leach AR. (2001) Prediction of Biological Activity for High-Throughput Screening Using Binary Kernel Discrimination. J. Chem. Inf. Comp. Set. 41 1295-1300. 

Historically, for every 100 compounds screened for pharmacologic activity in animal models, only one has the necessary biological activity and safety for evaluation in humans. Of those compounds tested in humans, only 1 in 40 is successfully brought to the marketplace. This poor rate of success has been attributed to a number of factors, a primary one being that animals are not truly predictive of biological activity or safety in humans. The problem, however, may be that insufficient time and resources are put into characterizing a lead... 

The results of the present study suggest the applicability of theoretical calculations such as frontier molecular orbital, dipole moments and AAHf in the prediction of biological activity of phenothiazines, benzo[a]phenothiazines, and benz[c] acridines. 

The third reason for the limited success of diversity-based strategies is the low baseline probability for bioactivity, with hit rates of 0.1% as the typical order of magnitude. In the case where a clustering was highly predictive of biologic activity, with active clusters that show a 100-fold enrichment of active compounds, this would still indicate that, on average, only 10% of compounds in any given active cluster are active. If the cluster was sampled with only one compound, the probability that the cluster is identified correctly as active would only be 10%. This theory is illustrated qualitatively in Fig. 1. 

PASS —prediction of biological activity spectra—compares a test structure with those in a database of about 45,000 structures with known activity/toxicity, using topological descriptors and probability calculations (125). 

Description Prediction of biological activities by substructure descriptors. 

Judson, P.N. (1992a). QSAR and Expert Systems in the Prediction of Biological Activity. Pes-tic.ScL, 36,155-1W. 

Rouvray, D.H. (1986b). The Prediction of Biological Activity Using Molecular Connectivity Indices. Acta PharmJugosl.,36, 239-252. 

MATRIX, a new algorithm for the prediction of biological activity of organic molecules based on multidimensional analysis of physicochemical descriptors of modern drugs 02ZOR1618. 

FIGURE 23.1 Structure—activity and structure—property relationships using data modeling techniques may provide die basis for understanding and prediction of biological activity and physicochemical features. 

The pseudo-receptor concept has been applied in recent years to analyze crucial ligand-receptor interaction sites and to establish 3D-QS ARs for the prediction of biological activities of ligands. A variety of application studies have shown that the pseudo-receptor concept is a versatile tool to establish 3D-QSAR models, often better in their predictive behavior compared to results obtained from classical 3D-QSAR approaches (e.g. CoMFA). Several application studies have been published which have shown the value but also the limitations of this approach. ... 

FIGURE 6.32 Base structure for a set of 44 progestagen derivatives used for training a CPG neural network for the prediction of biological activity. 

Bursi, R., et al.. Comparative Spectra Analysis (CoSA) Spectra as Three-Dimensional Molecular Descriptors for the Prediction of Biological Activities, J. Chem. Inf. Corn-put. Sciences, 39, 861, 1999. 

Deductive Inference Module Provides prediction of biological activity based on the rules stored in the knowledge base. 

Besides the statistical prediction of biological activity based on the rules, the deductive inference module provides explanation of predictions. It produces a three-dimensional graphic display of biophores found in the analyzed structures and allows superimposition of compounds with the same biophore. Visualization is performed via integration with the Insight II environment from Accekys Inc. [21], a molecular modeling environment with a powerful graphical interface. 

Bursi, R., Dao, T, van Wijk, T, de Gooyer, M., Kellenbach, E. and Verwer, P. (1999) Comparative spectra analysis (CoSA) spectra as three-dimensional molecular descriptors for the prediction of biological activities./. Chem. Inf. Comput. Sci., 39, 861-867. 

Filimonov, D.A. and Poroikov, V.V. (1996) PASS computerized prediction of biological activity spectra for chemical substances, in Bioactive Compound Design Possibilities for Industrial Use (eds M.G. Ford, R. Greenwood, G.T. Brooks, and R. Franke,), Bios Scientific Publishers, Portsmouth, UK, pp. 47-56. 

Optimising the EVA descriptor for prediction of biological activity. Org. Biomol. Chem., 2,3301-3311. 

Harper, G., Bradshaw, J., Gittins, J.C., Green, D.V.S. and Leach, A.R. (2001) Prediction of biological activity for high-throughput screening using binary kernel discrimination./. Chem. Inf. Comput. Sci., 41, 1295-1300. 

Judson, P.N. (1992a) QSAR and expert systems in the prediction of biological activity. Pestic. Sci., 36, 155-160. 

---